Introduction to applied statistics using R

First Meeting

• Introduction and logistics
  • Necessary software
  • Assessment, assignments, labs, etc.
• Intro to R

• What's the most effective way of collecting and analyzing data,
• Will cover key principles of statistical data analysis,
• Will get acquainted with key principles of working with R

• R;
• R-studio;
• A text editor;

• We are going to use TSU's E-Learning system for class management, therefore, you will need to get TSU email address.
• Instead of Facebook group, we will use Slack chat, where you'll be able to ask me questions, send feedback, share code & screenshots, etc.
  • access it here: itasur.slack.com
• Meanwhile, we will use my old website for distributing class materials: https://www.sichinava.ge/introstatsr/

Lecture

Lecture

Laboratory practice

Lecture

Lab

Literature

• Key text: Imai, K. (2015): Quantitative Social Science

Lecture

Lab

Literature

Assignments

• During the lab sessions, we will give you lab assignments which you have to submit throughout subsequent week through Dropbox link

• Attendance (5 points, detailed breakdown in syllabus)
• Assignments (10 lab exercises * 4 points each = 40 points)
  • It's important that I'm getting your lab assignments on time. Therefore, for each late week 1 point will be deducted from your mark. Say, you submitted one week later, then you'll get maximum 3 points instead of 3
• Midterm exam (open book, take home assignment)
• Final exam (conference paper)

• Replicable research: producing exactly the same results that were published (say, by running the same experiment)
• Reproducible research: ability to mirror the workflow of a particular study (by using the same dataset and running the same code)

Source: Marwick, B. (2014): Reproducible Research: A primer for the social sciences

• Standards (King, 1995):

The replication standard holds that sufficient information exists with which to understand, evaluate, and build upon a prior work if a third party can replicate the results without any additional information from the author

• Data reposotories (Dataverse @ Harvard, Figshare, Open Science Framework)
• Version control (Git, Github, Bitbucket)
• Preregistration (Open Science Framework, Center for Open Science, EGAP)

Source: Marwick, B. (2014): Reproducible Research: A primer for the social sciences

However…

The lexicon of reproducibility to date has been multifarious and ill-defined. The causes of and remedies for what is called poor reproducibility, in any scientific field, require a clear specification of the kind of reproducibility being discussed (methods, results, or inferences), a proper understanding of how it affects knowledge claims, scientific investigation of its causes, and an improved understanding of the limitations of statistical significance as a criterion for claims

Goodman, S., Fanelli, D., Ioannidis, J. (2016): What does research reproducibility mean?

• Verification and reliability;
• Transparency
• Efficiency
• Flexibility

Source: Marwick, B. (2014): Reproducible Research: A primer for the social sciences

😶 We enter data in Excel or SPSS (a.k.a., მონაცემთა ბადის მომზადება) 🤨 We clean data in Excel 😯 We do some data analysis in SPSS 😴 And copy and paste results in Word

All these can be organized at one place, so we can always go back and re-run our analysis

• Replication is useful not only for scientists but practitioners. A client might be interested how did you come up with your results
• Think about constant kerfuffle around public opinion polls in Georgia
• All prestigious (and not so prestigious) political communications and political science journals now ask about complete datasets and replication materials to publish any article
• It is just good practice

• Introduction to R
• R Graphic User Interface: R-Studio
• Key elements of R-Studio
• R-markdown documents
• Literate programming, replicable social research and the best practice for organizing your projects

• A programming language
• Freely distributed
  • GPL2-GPL3 license
• Available for all operational systems (Windows, Mac, Linux, etc…)

• For a long time, the lack of the GUI has been labelled as a main drawback for R users

• R-Studio does this job very well.

### General help:
help.start()

### Help on a particular function
help(lm) ## or
?lm

### Show me an example!
example(lm)


### Show me a vignette for a funtion/library
vignette("ggplot2-specs")

getwd()

setwd("D:/Dropbox/R/My awesome research")

or

setwd("D:\\Dropbox\\R\\My awesome research")

or

setwd('D:\\Dropbox\\R\\My awesome research')

install.packages("AwesomeLibrary")

library("AwesomeLibrary")

source("MyAwesomeRegression.R")

• Transparency
• Support
• Organize your code as modules
• Portability

1. Brief description of your study
2. Installing and loading all necessary libraries
3. Setting the absolute path from the beginning and then working using relative paths
4. Organizing your code in sections
5. Functions
6. Concise naming standards for your variables
7. Concise organization of files, folders and subfolders